[bigshot]

The MQA stance is that the original is before the final ADC, not the PCM.

So they're sourcing everything directly from analogue tape?! I flat out don't believe that... especially with modern recordings that were recorded DDD that have been digitized from the very first step of recording!

I think they take the 24/96 file the record label hands them. THAT'S as "authentic" a master as you can get, and it's no different than iTunes.

 

[pinnahertz]

So they're sourcing everything directly from analogue tape?! I flat out don't believe that... especially with modern recordings that were recorded DDD that have been digitized from the very first step of recording!

No, c'mon now, that's just crazy talk. Some stuff is analog, some/most is digits. How could they source everything from analog tape unless they made an analog tape copy of the digital master first? And that would just be dumb.

 

[bigshot]

How do you go back to the source before you convert it to digital without the master being analogue? Or perhaps a time machine. Unless I'm missing something, this makes no sense.

 

[jagwap]

No, c'mon now, that's just crazy talk. Some stuff is analog, some/most is digits. How could they source everything from analog tape unless they made an analog tape copy of the digital master first? And that would just be dumb.

I didn't say that and neither did they. I think some of you nay-sayers are deliberately choosing which bits of the information to remember to suit your arguments. They say they will try to use the best master, digital, analogue, shellac, what ever...

They claim they can model an post correct some of the issues of the last ADC used to go form analogue to digital. I assume this is for analogue master tape.

There is a technique that when you run a signal through DSP backwards, you can filter a signal the same as forwards, but if you match the forwards filtering (maybe the analogue bit), you can do it without altering the phase or the original. I guess they are trying something like this as I've said before.

 

[bigshot]

I guess I don't understand what they are claiming. They're saying that they're filtering the master quality digital file to correct for errors created the last time it went through analogue to digital conversion? It's kind of like pressing the "undo" button and removing the error and turning it back into the original pristine analogue signal again?

What if the signal had been processed in the digital domain several times since it went through that conversion? Wouldn't that processing muddy the waters and make it impossible to simply undo whatever problem the theoretical error caused? I can't imagine any master delivered by a record company to be a raw transfer directly off an analogue tape master. Every distribution master would have undergone some sort of mastering or restoration or perhaps sweetening since it was digitized.

The only way I can imagine that they would be able to know for sure that a master is a direct transfer with no modification is to do all of the digitizing and mastering themselves. Then they would have control of the process instead of trying to guess what had been done with the signal between the original analogue to digital conversion and the delivery of the master quality file to MQA. Re-digitizing and remastering everything would be reinventing the wheel on a massive scale with the size of the library they are working with. That just wouldn't be practical.

Also, if there really is some sort of "smearing" that they are able to correct by means of a backwards filter, that means that every single digital copy that exists- from CDs to SACDs to blu-ray audio to the digital masters themselves- all have this smearing. No one in the history of digital audio has ever heard music that has been recorded digitally without smearing. And logically, that means that the MQA file isn't just a faithful reproduction of the master, like Redbook, high bitrate lossy and HD audio. If their smear correction is audible, that means a streamed MQA file sounds *better* than the 24/96 digital master itself.

Why aren't recording studios taking advantage of this great technology? All they would have to do is apply the filter along with the analogue to digital conversion and it would improve the sound of the digitization! Why is MQA selling this to consumers listening to streamed audio instead of selling it to Digidesign and all the other companies that manufacture digital audio production hardware and software?

But of course, if you sell technology to sound engineers, you're going to have to prove with controlled testing that it's actually beneficial and explain how it works. I doubt they are willing to do that.

This is why I'm skeptical of the whole format. They make general statements about "smearing" and "authenticating" masters, but anyone who has any kind of experience in this stuff at all can see the outlines of the mirrors amid all the smoke. All of this stuff is in the hands of the label. By the time MQA gets handed the master, they can only work with what they're given... the same way Amazon and the iTunes store and Spotify and every other streaming service does. If there really is a technological breakthrough here, they should be selling it to audio production, not streaming.

 

[jagwap]

I guess I don't understand what they are claiming. They're saying that they're filtering the master quality digital file to correct for errors created the last time it went through analogue to digital conversion? It's kind of like pressing the "undo" button and removing the error and turning it back into the original pristine analogue signal again?

What if the signal had been processed in the digital domain several times since it went through that conversion? Wouldn't that processing muddy the waters and make it impossible to simply undo whatever problem the theoretical error caused? I can't imagine any master delivered by a record company to be a raw transfer directly off an analogue tape master. Every distribution master would have undergone some sort of mastering or restoration or perhaps sweetening since it was digitized.

The only way I can imagine that they would be able to know for sure that a master is a direct transfer with no modification is to do all of the digitizing and mastering themselves. Then they would have control of the process instead of trying to guess what had been done with the signal between the original analogue to digital conversion and the delivery of the master quality file to MQA. Re-digitizing and remastering everything would be reinventing the wheel on a massive scale with the size of the library they are working with. That just wouldn't be practical.

Also, if there really is some sort of "smearing" that they are able to correct by means of a backwards filter, that means that every single digital copy that exists- from CDs to SACDs to blu-ray audio to the digital masters themselves- all have this smearing. No one in the history of digital audio has ever heard music that has been recorded digitally without smearing. And logically, that means that the MQA file isn't just a faithful reproduction of the master, like Redbook, high bitrate lossy and HD audio. If their smear correction is audible, that means a streamed MQA file sounds *better* than the 24/96 digital master itself.

Why aren't recording studios taking advantage of this great technology? All they would have to do is apply the filter along with the analogue to digital conversion and it would improve the sound of the digitization! Why is MQA selling this to consumers listening to streamed audio instead of selling it to Digidesign and all the other companies that manufacture digital audio production hardware and software?

But of course, if you sell technology to sound engineers, you're going to have to prove with controlled testing that it's actually beneficial and explain how it works. I doubt they are willing to do that.

This is why I'm skeptical of the whole format. They make general statements about "smearing" and "authenticating" masters, but anyone who has any kind of experience in this stuff at all can see the outlines of the mirrors amid all the smoke. All of this stuff is in the hands of the label. By the time MQA gets handed the master, they can only work with what they're given... the same way Amazon and the iTunes store and Spotify and every other streaming service does. If there really is a technological breakthrough here, they should be selling it to audio production, not streaming.

This is Why they are offering streaming and other end to end playback, as I think they need to have the filtering compensation with as much known phase and timing behaiour in the ADC and DAC as possible to minimise the timing error. (Please note I am not stating this effect exists, this is just my interpretation of what I think they may be trying to achieve). You cannot pass the whole audio signal backwards through DSP while you are working in it. (I am still looking for a windowing techique rhat allows this, but my DSP colleagues haven't found one yet. I need to widen the search). Perhaps if this all takes off each CODEC in the recording chain could add a stamp of its behaviour in the AES EBU user bits to be used later by the MQA encode, but until then they are doing what they can.

As to sweeting masters, I think this is what the Authenticated bit intends to avoid. They want the master before the sweeeting, or the one after the speed correction if it it is more accurate, etc... They have said they can only do their best, and of course it needs the cooperation of the lables and artist to work well.

At least we are now discussing itb rather than the previous "it cannot do anything good" from others here.

 

[bigshot]

How do they plan to achieve end to end? Are they going to be pulling original tape masters and mastering them themselves? What if all that is available is a 24/96 distribution master of something that was recorded analogue? Is that going to not be authenticated for MQA because it's been processed? Who is going to pay for all this work? The labels themselves? Does MQA's technology have nothing to offer to albums like Donald Fagen's The Nightfly which was recorded digitally from end to end at 16/44.1?

 

[jagwap]

How do they plan to achieve end to end? Are they going to be pulling original tape masters and mastering them themselves? What if all that is available is a 24/96 distribution master of something that was recorded analogue? Is that going to not be authenticated for MQA because it's been processed? Who is going to pay for all this work? The labels themselves? Does MQA's technology have nothing to offer to albums like Donald Fagen's The Nightfly which was recorded digitally from end to end at 16/44.1?

One case: digital master where the ADC is known - add ADC correction to the digital file and tell the MQA DAC which filter to use.

Next case: Digital master is not good, analogue master is available - remaster with an MQA ADC.

They have said multiple times they have many approaches, some better than others. They plan to do as much as they can to minimise the effect of recording chain on these "temporal effects". They also say the music comes first, which is commendable, but relient on the labels.

The Nightfly MQA version has just been released. Sounds good, and better than the non MQA version on Tidal. Not scientific, but it is either a better master, less compressed or MQA special sauce, or a combination of the above. But yes, if there is no ADC at this stage there is nothing to correct. As others have said, it will be hard to correct ADCs further up the chain, particularly retrospectively.

 

[bigshot]

I don't believe that there is any way to know what ADC was used and how many times the signal has been processed since it passed through the ADC. In a recording session every single track goes through an ADC separately. If smearing is going on there are layers and layers of smearing stacked on top of each other in the final mix. The only way what you're saying is possible is if they are sourcing a tape master that was recorded and mixed analogue. I don't believe MQA has access to those. I think they're working from digital distribution masters that have undergone all kinds of digital filtering. If you are describing it properly, I think it is a fabrication for the purposes of sales pitch.

If The Nightfly sounds better than other releases, it isn't because of anything MQA did. It can't contain information above 22kHz because it was recorded and mixed at 16/44.1. Their magic mojo unsmear filter can't be credited because it only passed throug an ADC once as it was being recorded and it was subsequently mixed and mastered. It cant be a better generation master because digital isn't subject to generation loss. The only reason it could sound better is because...

Wait for it...

They sweetened it and it isn't the original mix and mastering that Fagen approved. Do you think it might actually sound exactly the same as every other format? (Giving you a chance to think about it and back out from this line of thought.)

 

[jagwap]

Your opinion and you are entitled to it. But there is no eveidence either way yet.

I'm just stating what I understand to be their position and what I think that may mean as a counter point

Material before digital will go through one ADC when first mastered to digital for example when being released for CD. Then it may be clearer what happened, particularly as there were limited choices back then what could be used. I imagine it is not hard to work out the ADC then, and it would not be "state of the art" compared to now.

By the way, I just checked The Nighfly. When played back with MQA it is 44.1kHz as it should be. Not upsampled or falsely represented as 192kHz. I can't see the bit depth yet.

 

[bigshot]

It isn't my opinion. Donald Fagan's The Nightly would be impossible to improve given your description of what MQA is doing. Tell me why it could sound better on MQA.

 

[jagwap]

It isn't my opinion. Donald Fagan's The Nightly would be impossible to improve given your description of what MQA is doing. Tell me why it could sound better on MQA.

Because the DAC also apparently has time smearingb and pre-compensating in the digital file before playback allows phase free correction as I described above.

 

[pinnahertz]

One case: digital master where the ADC is known - add ADC correction to the digital file and tell the MQA DAC which filter to use.

There's an ADC on every input of a multitrack/multi input system. They aren't always identical, sometimes not even the same manufacturer. And sometimes they are. Then there's multiple trips through ADC/DAC/ADC/DAC for processing. The instances where it was one, known ADC only are miniscule.

Next case: Digital master is not good, analogue master is available - remaster with an MQA ADC.

Which fixes nothing. Anyone who's ever run a square wave through a tape recorder knows...yep, it has it's own version and degree of "temporal blurring". Also uncorrectable.

They have said multiple times they have many approaches, some better than others. They plan to do as much as they can to minimise the effect of recording chain on these "temporal effects". They also say the music comes first, which is commendable, but relient on the labels.

From the MQA web site, their #1 tenant and major definite purpose: "MQA captures 100% of the original studio performance." Direct quote, and also direct lie.

The Nightfly MQA version has just been released. Sounds good, and better than the non MQA version on Tidal. Not scientific, but it is either a better master, less compressed or MQA special sauce, or a combination of the above. But yes, if there is no ADC at this stage there is nothing to correct.

Well sure there was! It's a digital recording now, isn't it? How'd that happen? My money is on an ADC.

As others have said, it will be hard to correct ADCs further up the chain, particularly retrospectively.

No, it's hard to correct for a single ADC, it's impossible to correct for an unknown ADC or unknown multiple ADCs. And all of this is retrospective at this point, no new material of any consequence.

Look, this has all be done and argued to death. In know I've typed all of this or equiv. multiple times. There's only one way to end it: unassailable proof. And that would take MQA's co-operation, financial resources, and time. It's the infernal triangle, take away any one, you don't have it.

 

[Strangelove424]

If The Nightfly sounds better than other releases, it isn't because of anything MQA did. It can't contain information above 22kHz because it was recorded and mixed at 16/44.1. Their magic mojo unsmear filter can't be credited because it only passed through an ADC once as it was being recorded and it was subsequently mixed and mastered. It cant be a better generation master because digital isn't subject to generation loss. The only reason it could sound better is because...

Wait for it...

They sweetened it and it isn't the original mix and mastering that Fagen approved. Do you think it might actually sound exactly the same as every other format? (Giving you a chance to think about it and back out from this line of thought.)

Mastering their own version from some renegade studio, without artist supervision! I can't imagine anything less authorized than that! The simpler, far more believable option is that MQA is nothing but an inaudible 22khz+ stream parallel to the actual music. But what a logical sword to fall on! Either the music sounds pure and MQA can't add or take away anything, or it sounds better with MQA but it isn't pure! How can MQA satisfy audiophile demand for something to be improved and yet original and pure at the same time? It's a logical contradiction, but logical contradictions never get in the way of marketing. (or audiophiles... a pair made in heaven)

 

[jagwap]

It's like being under a Danish bridge: full of trolls.

But then you probably see me the same way...

There's an ADC on every input of a multitrack/multi input system. They aren't always identical, sometimes not even the same manufacturer. And sometimes they are. Then there's multiple trips through ADC/DAC/ADC/DAC for processing. The instances where it was one, known ADC only are miniscule.
Which fixes nothing. Anyone who's ever run a square wave through a tape recorder knows...yep, it has it's own version and degree of "temporal blurring". Also uncorrectable.

Not on ALL recordings. Music was recorded before digital. But it does raise issues.

Also your square wave point is not adding to the discussion. If all we have is the analogue master before digital the PCM and the MQA will have the same problems.

From the MQA web site, their #1 tenant and major definite purpose: "MQA captures 100% of the original studio performance." Direct quote, and also direct lie.

Marketing has always lied. Get over it. "Perfect sound forever". We discussed that before. Forever is BS. Here we want the science, so lets delve into that.

Well sure there was! It's a digital recording now, isn't it? How'd that happen? My money is on an ADC.

No, it's hard to correct for a single ADC, it's impossible to correct for an unknown ADC or unknown multiple ADCs. And all of this is retrospective at this point, no new material of any consequence.

Look, this has all be done and argued to death. In know I've typed all of this or equiv. multiple times. There's only one way to end it: unassailable proof. And that would take MQA's co-operation, financial resources, and time. It's the infernal triangle, take away any one, you don't have it.

True. Argued to death and neither of us are proved right. But you are still arguing, like me. So what's your point?

Mastering their own version from some renegade studio, without artist supervision! I can't imagine anything less authorized than that!

No one is saying that. Not sure why you are. Can you enlighen us?

The simpler, far more believable option is that MQA is nothing but an inaudible 22khz+ stream parallel to the actual music. But what a logical sword to fall on! Either the music sounds pure and MQA can't add or take away anything, or it sounds better with MQA but it isn't pure!

Not according to MQA. They are stating that the ADC and DAC is not as good at representing the analogue as it could be, so this "true" you are defining as the digital master is not their definition. There are many other processes they cannot improve but they are trying for this one.

How can MQA satisfy audiophile demand for something to be improved and yet original and pure at the same time? It's a logical contradiction, but logical contradictions never get in the way of marketing. (or audiophiles... a pair made in heaven)

Not their view I think.